Homing offsets for slaved axes

[Matty Zee]

Hi,

I've posted this question over at Warp9 as I'm using a Smoothstepper but i think its more of a Mach3 question.

I'm trying to use homing to square my gantry. I have a limit switch on both side of the gantry. Problem is, my switches on either side of the gantry aren't aligned perfectly. I can't easily adjust the position of my sensors. I was hoping to be to provide a different home offset for either side of the gantry. If i set different homing offsets for the two axes, the slaved axis offset is ignored.

The only way i can imagine i can do this is to unslave the two axis, home simultaneously and reslave. Can i do this? I've tried but it seems to go all wrong and axes end up mot re-slaving properly.

Cheers,
Matty

[mhasting2004]

Hi Matty

Not a fix but you should consider the fact that you will be racking your gantry every time you home even if it is possible to have different offsets.... might make the effort in making the home switches adjustible.

For a hardware fix and something I will do when I finally put servos on my machine how about this... use the home sensor as a gate/load signal to an up/down counter with a preset number (selectable via dip switch) and have the terminal count signal be latched as the new "repositioned" home signal. Clock signal for the counter is from the encoder on the axis.

Clear as mud? This works very well and I used to use this to reposition the index pulse on high end imagesetters I helped design.

Cheers
Mark

[mhasting2004]

Just had another thought.. if you don't have encoders on your stepper or a servo setup the you could possibly use the step input to your drives as the clock signal. Note also with a single 8 bit counter you have up to 256 count delay .. cascade another one and you have heaps (64k) of possible delay. You would only need to do this to the side that has the sensor that is normally triggered first.

You may also need to put a counter in for the deactivate position as I believe that is where Mach sets its zero .. plus a step off amont... which I'm currently trying to find out how to set.

[Matty Zee]

Thanks Mark. Will consider that if i put servos on my machine but its not in the plans at the moment. (need to let it sink in though, still a bit muddy) ;-)

My gantry racks a fair bit as it is with the way Mach3 does gantry homing. I'm using inductive sensors and its hard to position them accurately. Installed height has a big impact. I have them within about 0.2mm but i think i'm at the point where i'll be chasing my tail. Unfortunately the Mechmate design didn't allow well for this kind of thing.

I just want to be able to measure the offset with calipers and input the value into Mach3. Didn't think that was asking too much. Not sure if Mach3 doesn't do it or if the SS doesn't.

[mhasting2004]

Same issue for me.. inductive sensor etc. I had plans of putting a rotary encoder on the arse end of my X axis steppers as they are dual shaft but not doe it yet. and I have a set of servos on the shelf that I really should try out. Somehow need a fixed in space index pulse off the drive... because as you say if the home sensor is not repeatable then its no use having an offset as it would need to constantly change. Also if its not repeatable its not a good choice as our initial index pulse... so that negates a bit of what i was suggesting before.

BTW I agree 100% about the homing routine.. not a big fan in its current form as it does rack the gantry slightly the way it works at the moment. Posted a request to get a change ahiile back (SS issue) but at the time backlash comp was the big todo item.. not sure if Greg has had a chance to revisit this yet. If you find out where the step off distance is hidden please give me a heads up.

As to Mud...
To clarify.. or at least attempt:
Say you have an encoder and and an index pulse and you want to have the index at a precise position in the rotation. You have two choices.. mechanically rotate and fix the encoder to the shaft in precisely the correct position or recreate a new index pulse "n" encoder pulses later. Using a counter circuit allows you to do this.
Using the actual prox as the enable or gate signal to make the delayed index valid.